> > The equal(%e,2.718281828459045) fact is bogus.
The simple solution is assume(%e > 2.718, %e < 2.719) (to appropriate
precision, of course).
The problem is that, though Compar is smart enough to make simple
deductions from this like is(%e*100 < 300), it can't make more
complicated ones like is(%e^2 < 10). In fact, is doesn't even use
'equal' facts in cases like this: assume(equal(s2,1.14)), sign(s^2-2)
=> pnz
I'm not sure I understand why using the numerval is any better than
using the Compar database; doesn't it also have an approximate value?
-s